Design Optimization of Railway Steel Bridges Utilizing High Frequency Mechanical Impact Treatment

Abstract

Railway bridges are highly susceptible to fatigue, which typically governs their de sign. By applying High Frequency Mechanical Impact (HFMI) treatment to enhance fatigue strength, new opportunities arise to effectively utilize Higher Strength Steel (HSS) without increasing dimensions for the cross-section. To evaluate this poten tial, a Parametric Design Model (PDM) was developed in MATLAB and combined with a Genetic Algorithm (GA) to optimize simply supported, single-track, twin I-girder railway bridges (10–40 m spans) across steel grades S355, S460, and S690. The GA optimizes the cross-sectional geometry to minimize total initial produc tion costs taking encompassing material, welding, and painting in to account while strictly verifying the designs against Fatigue Limit State (FLS), Ultimate Limit State (ULS), and Serviceability Limit State (SLS) criteria. This thesis then investigates whether it is profitable, and in which combinations of HFMI and HSS, the potential for material and cost savings is greatest. The re sults demonstrate that the most significant cost savings are achieved using HFMI in combination with steel grade S355. While steel grades S460 and S690 provide some benefits compared to the As-Welded (AW) case for shorter span lengths of 10 m, its economic viability is severely restricted for longer spans. A key finding regarding the failure modes is that while FLS strictly governs the AW that is used as the reference designs case, the application of HFMI shifts the governing failure mode to ULS for shorter spans, and entirely to SLS deflection for spans over 20 m. To get representative results the optimization was done considering total investment cost. Conversely, data was analysed comparing both impact on CO2-emissions and total Life Cycle Cost (LCC). These comparisons implied some even greater potential for design optimization utilizing HFMI if full life cycle were to be considered. In addition, an easier comparisons of fatigue design methods was made, which indicated that the positive benefits from HFMI is method independent.